The “scientific method” itself is flawed. As soon as you put a human in to observe anything, a study is rife with issues. This is why they do double-blind studies (take the observer’s control away as much as possible) and meta studies (shove all related studies together and data-mine, to reduce human error that is higher in smaller batches). Even then, it’s not perfect.
It’s a part of why I’ve long looked at redactions and watchdog groups over studies and have accepted that bad faith actors who screw up one study are likely to be back in another study next year, without any effort to keep track of their involvement from year to year. So while I do look at outcomes of studies, I’m often skeptical of the conclusions, as it is.
When dealing with behavioral studies which produce the worst in mankind, those outcomes are just a hair more trustworthy because the observer’s evil is naturally a part of the system which created the outcome. The confusion is solely in trusting the advocacy of effectiveness for treatments–becauee no, you don’t trust things that don’t come from care for people in safe treatment of them. But the results still have solid points. It’s what makes it scary in the first place.
For example: we know from studies of nervous reactions that people really only have the bodily functions to panic for about half an hour. (At least this was the conclusion 15ish years ago that may have changed some.) That means that it doesn’t matter if we get through a study that puts people through sheer hell for an hour or so or a more safe study that takes people who are naturally panicking and try to calm them down while documenting the baseline rate of panic. Either way, the trends will show that half an hour is the norm for life-endangering panic.
It also means that people who are going through panic attack triggers may over come phobias by stubbornly outlasting the reaction In fact some of the humane treatments are exposure while reminding people that they can outlast their triggers…but they give the power to walk away to the person who is suffering. I’ve watched documentaries on therapy sessions like that.
It’s also why I choose to ride out some events myself. Can’t let a phobia get to me while I’m driving–that’s a good way to die. The ethics of such studies play no part in stopping me.
Ethics, instead, has far more to do with the value I place on life. Humans have lives of merit. Even if animals are of “lesser value”, my merit as a human demands I don’t overly mistreat animals, either. Some outcomes simply lack merit because they do not treat the subject with respect and has nothing to do with outcomes.